The Push-Up Method for Reading Speed (And How Much of the Gain Sticks)

Your Normal Pace Isn't a Limit, It's a Habit

You sit down with a book, read for a minute, and settle into a tempo that feels neither rushed nor slow. It's just your pace.

Who picked it?

You never negotiated it. It set itself, back in primary school, when you read out loud and the speed of your mouth was the speed of your eye. You've changed books many times since, but nobody ever went back to check whether that tempo still matched what you're capable of.

That's where the push-up method starts. It doesn't ask you to read fast all the time. It asks you to step out of your usual tempo for two or three minutes, deliberately too fast, and then step back into it. When you return, the pace that used to feel normal feels sluggish.

The sensation is real, and it has an explanation. What's left of it three weeks later is a different question, and that's where the people who measure part ways from the people who sell.

The Overload Principle, Borrowed From the Gym

In strength training, the rule that does all the work is called progressive overload. For a muscle to grow, it has to work slightly past what feels comfortable today. Not far past — just enough to stop being routine. Then comes rest, and the adaptation happens during the rest, not the effort.

The reading push-up is a direct transposition of that idea, into a session with three parts.

The first is the baseline round: you read for two minutes at your usual pace and note how much you covered. The second is the forced round: three minutes in which you deliberately push faster than you can comfortably track, and comprehension drops. The third is the return round: you go back to your normal tempo, without forcing anything, and pay attention to how it feels.

The whole method lives in that third round. The first two are just the setup for it.

Why the Old Pace Feels Slower Afterward

You've already felt this effect, in another context. You come off the highway after an hour at 130 km/h, merge into city traffic at 50, and it feels like you're standing still. The city's speed hasn't changed from yesterday. The reference point in your head has.

Something similar happens with reading, plus a part that matters more.

The eye doesn't glide along the line. It stops, jumps, stops again. Rayner pulled together the measurements from dozens of studies in 1998: fixations last around 200 to 250 milliseconds, each jump covers seven or eight letters, and somewhere between 10 and 15% of eye movements are backward jumps over text already read.

Not all of those backward jumps are necessary. Some repair a sentence that genuinely got tangled syntactically. Some are reflex, a safety net thrown out by a reader who doesn't trust that they caught the word the first time. The same goes for time spent on a word: part of it is recognition, part of it is the inner voice finishing the pronunciation of something the eye already knew.

The forced round doesn't hand you new capacity. It cuts off access to the safety net. Three minutes in which you simply don't have time to double back, and you discover the text still holds together without the backward jumps.

When you return, two things overlap. One is contrast, like coming off the highway, and it fades within a few minutes. The other is a genuinely cleaner pass, with less dead weight. These are different phenomena, and the fact that they feel identical is exactly why the sensation right after the drill isn't allowed to be your measurement.

Where the Gym Analogy Breaks Down

In the gym, overload works because the muscle actually adapts to the load. In reading, the thing holding you back isn't a muscle.

What holds you back, for the most part, is how fast you recognize words. Kuperman and Van Dyke tested a whole battery of verbal and cognitive skills in 2011: word recognition speed predicted reading speed better than anything else on the list. Not eye strength. Not the width of your visual span.

And word recognition doesn't get trained in three minutes of forcing the pace. It grows slowly, through volume, because every encounter with a word makes it a little cheaper to process the next time.

That's where the prediction the method never tells you comes from, even though the data backs it up: the first few weeks climb nicely, then the curve flattens. It climbs because you have dead weight to cut, and dead weight cuts fast. It flattens once the dead weight is gone and what's left is the real ceiling.

A controlled study from 2023 shows roughly how much "nicely" is worth. Klimovich, Tiffin-Richards, and Richter compared a commercial speed-reading program against a group that got only an hour of explanation about how reading works. Both groups gained a few dozen words per minute over the control group, and the difference between them wasn't statistically significant.

That sentence is worth rereading. An hour of explanation produced roughly what a full training program did. That doesn't make the exercise pointless, but it shifts the explanation: a good chunk of the gain in the first few weeks comes from understanding what you were doing wrong, not from a new capacity installed through repetition.

Then there's the price. The field synthesis by Rayner, Schotter, Masson, Potter, and Treiman from 2016 doesn't soften this point: the trade-off between speed and comprehension is real and can't be engineered away. Push the pace past what you can process, and you pay for it in comprehension. Their practical conclusion is that a normal reading pace is good enough for most purposes.

So the push-up recovers waste. It doesn't move your ceiling.

The Protocol, Concretely

Eight minutes, three times a week, three weeks. That's it.

The text. A book you'd read anyway, in prose you're comfortable with. Not the textbook you're being examined on, not a technical text from a field you don't know. On hard material, what slows you down is word recognition, and that can't be forced.

Baseline round, 2 minutes. Your normal pace, nothing special. At the end, mark your spot and count the pages or lines. That's the unit of measurement for the session.

Forced round, 3 minutes. The target is fifty percent more text per minute than your baseline round. If you covered four pages in two minutes, you need to cover nine in three. You'll lose detail, and that's fine — the loss is part of the mechanism, not a sign you did something wrong.

To hold the pace you need a metronome, otherwise you slow down quietly without noticing. The simplest option is your finger or a card you move at a set speed, with your eye following it. The evidence that finger-guiding makes you a faster reader is thin — I looked at all of it and wrote up what I found — but here you're asking it for something different: not to accelerate you, just to stop you from slowing down.

Return round, 3 minutes. Back to your normal tempo. Don't try to hold the fast pace. This round has no target — it has an observation.

What you don't do. Don't run it daily until you're worn out, because the adaptation doesn't come from volume. Don't do it on the chapter your grade depends on. Don't turn it into a chase for a number.

How You Know Whether Anything Actually Grew

The feeling right after the return round is the worst measuring instrument in the drawer, precisely because it's partly contrast, and contrast fades on its own.

An honest measurement has three conditions. Take it cold, not right after the drill. Take it on text of comparable difficulty to your first measurement. And include comprehension, not just the stopwatch, because a speed with no comprehension percentage next to it says nothing.

What numbers are realistic for an adult, why there's no published norm for this, and what an honest measurement looks like — I wrote about that separately. Here, the rule is enough: measure today, measure again in three weeks, read the two numbers together.

A reasonable expectation is a gain of a few dozen words per minute over three or four weeks, with no loss of comprehension. If you see a doubling, check what you actually did. Most likely you shifted from reading to skimming, which is a legitimate tool, but a different one.

What No Speed Drill Can Force

There's another part that never shows up on the stopwatch. You can get through a chapter 15% faster and retain nothing from it, in which case you've just bought yourself extra time to forget more efficiently.

Speed and retention train separately. The cheapest check takes ten seconds: at the end of every chapter, close the book and say out loud what was in it. Whatever doesn't come out wasn't there. I wrote at length about why rereading produces the feeling of knowing, while testing produces the memory.

The push-up method is good at what it's good at. It shows you that the tempo you thought was fixed is actually negotiable, and it clears out some of the waste. The rest — the part that grows slowly and doesn't show up in a single session — comes from how much you read, not how hard you force it.

→ Measure your reading speed and comprehension, free